https://peer.madiator.cloud/videos/watch/d7ba5597-fa44-4ed4-a007-e2e17f46213e NOTE: If you are using 5090s or B200s (Blackwell GPUs) they require a different template (the Blackwell template at runpod/comfyui:latest-5090. Check out this addendum for a quick update. https://www.youtube.com/watch?v=5-CIu2GWiDk Get started today at https://www.runpod.io Download option timestamps: 00:00 Overview of Template 02:09 FileBrowser (drag and drop) 03:43 Terminal - wget 04:07 Terminal - git clone 05:02 Terminal - CivitAI API download This video introduces Runpod's new official ComfyUI template, which includes the latest version of ComfyUI, KJ nodes for working with popular diffusion models, a Jupyter notebook for file management and terminal access, and File Browser for more efficient file transfers. Unlike previous official templates, this one-size-fits-all template works with any GPU (including 5090s and B200s) and intentionally doesn't include pre-installed models. This design choice prevents users from having to download unwanted models that slow startup times and take up unnecessary storage space. The video demonstrates four methods for getting models into your pod: using File Browser's drag-and-drop interface for smaller files like LoRAs, downloading single files from the internet using wget commands in the terminal, cloning entire GitHub or Hugging Face repositories with Git LFS for bulk downloads, and using Civitai's API to download models directly from their platform. The tutorial concludes with a practical demonstration showing how to run SDXL using the downloaded checkpoints and pre-built workflow templates, proving how quickly users can get up and running with their preferred models. Music Credit: Abandon Planet by HOME Tue, 21 Jul 2026 20:32:58 GMT https://validator.w3.org/feed/docs/rss2.html PeerTube - https://peer.madiator.cloud https://peer.madiator.cloud/lazy-static/avatars/e2927aea-9e0a-454c-9327-68ec87b4da11.png https://peer.madiator.cloud/videos/watch/d7ba5597-fa44-4ed4-a007-e2e17f46213e All rights reserved, unless otherwise specified in the terms specified at https://peer.madiator.cloud/about and potential licenses granted by each content's rightholder.